人工智能培训

搜索

机器学习论文:从长期短期记忆(LSTM)架构中的序列学习中提取知识(Knowledge extraction from the learning of sequ

[复制链接]
websun01 发表于 2019-12-9 13:48:22 | 显示全部楼层 |阅读模式
websun01 2019-12-9 13:48:22 205 0 显示全部楼层
机器学习论文:从长期短期记忆(LSTM)架构中的序列学习中提取知识(Knowledge extraction from the learning of sequences in a long short term  memory (LSTM) architecture)根据未知的生成自动机,我们介绍了一种从递归神经网络(长期短期记忆)中提取知识的通用方法,该知识已学会检测给定输入序列是否有效。基于隐藏状态的聚类,我们解释了如何构建和验证与基础(未知)自动机相对应的自动机,并允许预测给定序列是否有效。该方法在人工语法(Reber的语法变体)上以及在实际用例中均已阐明,但其基础语法是未知的。
We introduce a general method to extract knowledge from a recurrent neuralnetwork (Long Short Term Memory) that has learnt to detect if a given inputsequence is valid or not, according to an unknown generative automaton.Basedon the clustering of the hidden states, we explain how to build and validate anautomaton that corresponds to the underlying (unknown) automaton, and allows topredict if a given sequence is valid or not.The method is illustrated onartificial grammars (Reber's grammar variations) as well as on a real use-casewhose underlying grammar is unknown.机器学习论文:从长期短期记忆(LSTM)架构中的序列学习中提取知识(Knowledge extraction from the learning of sequences in a long short term  memory (LSTM) architecture)
URL地址:https://arxiv.org/abs/1912.03126     ----pdf下载地址:https://arxiv.org/pdf/1912.03126    ----机器学习论文:从长期短期记忆(LSTM)架构中的序列学习中提取知识(Knowledge extraction from the learning of sequences in a long short term  memory (LSTM) architecture)
回复

使用道具 举报

您需要登录后才可以回帖 登录 | 立即注册

本版积分规则 返回列表 发新帖

websun01当前离线
新手上路

查看:205 | 回复:0

快速回复 返回顶部 返回列表